There has been disclosed a technique for inferring a person's emotion from the person's voice, specifically from the volume of the voice, the waveform of the volume, the pitch of the voice, the phonetic of the voice, etc. (Patent Document 1). In addition, there has been disclosed another technique for expressing a human's emotion by three-dimensionally changing and displaying expression of a simulated model of a human's face created on 3D computer graphics on the basis of the emotion inferred by the technique according to Patent Document 1 or the like (Patent Document 2). By these techniques, a talking person's emotion can be read visually from only the person's voice.
Patent Document 1: WO 00/62279
Patent Document 2: JP-A-2002-298155